When to play "real bridge"

People are fond of enumerating the differences between IMPs (or total points) and matchpoints. True, there are lots of differences, although there are more similarities. Good bridge is good bridge and bad bridge is... well you know the answer.

This article is about the play, not the auction. Much of the difference between forms of scoring manifests itself in the auction. Once we get to the play, the differences are fewer. Oh, but don't they say that you should always try for overtricks at matchpoints? No, they don't.

In a team game, you are in what I call "real bridge" mode.  As developed by Vanderbilt, the idea behind contract bridge is that, once you're playing a particular contract, you must devote every effort to making (or defeating) it. You stop worrying about better contracts that you might have been in. And you take absolutely no risks for over (or under) tricks. If you're on defense and you have the setting trick, you cash it!

But, when you're playing matchpoints, it's tempting to go all out for overtricks all the time. That's not right. Here's where you try for overtricks: when you're in a contract that you judge most others will be in. In that case, you need to strive for the maximum possible number of tricks, even if that risks you going down. However, the majority of contracts are not like that. I'd like to talk about the contracts where you should not play for overtricks.

Basically, it's any time you're in an unusual contract. Let's take the most obvious example first: the contract was doubled. Once you're doubled, overtricks essentially no longer count. Even if you're only in, say, 1NTX not-vulnerable, if you make it, you will beat anyone undoubled making two, or three no-trump. However, if you think that a more normal contract is 3NT making three (and you forgot to redouble), then taking nine tricks will only score 380 (580 if vulnerable). You actually need to make 10 tricks to beat the others making nine tricks. But this is an unusual situation. Normally, the bonus for the insult and the "odd" tricks counting double will render one or two overtricks pure "gravy."

Another exception is when you're in a doubled contract and you're expecting to go down but are making a sacrifice. Now you have a precise target, which makes it a lot more like you are trying to "make" a "contract", albeit two or three tricks fewer than the nominal contract.

Other competitive situations are very much more like "real bridge." Imagine that you are defending 3S after your side competed to 3H. In other words, you pushed them into 3S. Nobody doubles so it might seem that this should fall into the "normal" contract category. Not so. You can't be sure what will happen at the other tables but you can be sure there will be some playing 2S their way and some playing 3H (maybe even 2H) your way. You might argue that, if 2H or 3H is making, and the opponents are vulnerable, you will need a two-trick set to restore equity. But you can't be sure that you actually had equity of 140. You surely didn't have 110 because the opponents can easily outbid you in 2S.

It's a cardinal rule of bridge that you must assume that you are in the right contract, whether declaring or defending. In other words, trust your bidding judgment. This is because, if we assume the converse, that we messed up in the auction, then it's really unlikely that we can recover our error in the play. Therefore, defending 3S is the proper place to be and there should be a big difference between -140 and +100. So, let's assume that you are offered the setting trick against 3S. DO NOT DUCK IT! If ducking gets you an extra trick for 200, then technically, you were in the wrong contract. You should have been defending 3SX.

But more often than not, a one-trick set will be enough: at least to stay average, which is what you should aim for on a hand like this.

Do I follow my own advice? Of course not! I ducked the setting trick just the other day thinking I could get 200 or at least I'd end up with the same 100. I was wrong. I ended up with a very poor-scoring -140. A bottom, in other words. So, I will plead the fifth: I refuse to incriminate myself by a hand diagram.

There are a few other situations where you should play "real bridge" in a matchpoint tournament. An example is when you bid a slam, especially if it's a hard-to-reach slam. Unless you're playing in an NABC pairs event, half the field won't bid even the most obvious slam. So, don't risk your slam for the overtrick that you think you might be able to get. You already won this board in the auction. Just sit back and collect.

Other strange contracts like 2NT, 5NT, a non-competitive 4 of a minor, even a 2C contract, are all sufficiently unusual that you should probably just ensure your contract.

An interesting hand at matchpoints

A hand came up in a BBO speedball which caused a dent in an otherwise decent session. First let's look at it from my (declarer's) point of view.

The opening lead is D9. There's really very little time in a speedball to ask (or consult) about lead conventions. East played the J and I won with the Ace. What do you think of your chances in this contract? Since West didn't lead a heart (presumably, he has the ace), we have a shot at making 12 tricks. Partner did the right thing however, in signing off in 4S with that wasted HK.

Still, that diamond lead is a bit awkward. My only entry to the dummy now is by ruffing the small diamond. But I can't now draw all the trumps before taking the club finesse. So, I took two rounds of trump and ruffed the diamond. Now, I took the club finesse. Good news: it won. Bad news: the trick was ruffed by West. Now, I had to lose to the CK too and just made my game. This was worth a paltry 30% of the board.

It turns out that the way to make five (or more) on this hand is to be either a) lucky; b) forgetful (or insanely optimistic); c) brilliant; d) playing double-dummy. Needless to say, I was none of those. With 128 tables in the event, you can imagine that there were all sorts of results, but the vast majority of the NS pairs were in 4S.

At one table which also received a diamond lead, declarer didn't worry about that losing diamond, drew trumps, took the club finesse a couple of times and conceded two red cards. At many other tables, declarer was fortunate to receive the lead of HA. Now, there's no excuse for not making six (although many such declarers still only made five).

I mentioned that you can make 5 playing double-dummy. How so? Well, since we know we have to lose a heart anyway, how about playing a heart up after drawing two rounds of trump (or playing the HK after ruffing the diamond)? Now, South has no choice but to draw any remaining trumps and try to drop a singleton CK. It won't work but at least he doesn't suffer a ruff.

Interesting hand.

Dilemma

Well, it seems that I have figured out a way to publish my blog again. It's not that I didn't want to. I have lots of material. It's just that I couldn't show you any actual hands because Google broke the Blogger code.

I tried to join BridgeBlogging (where Jeff Lehman publishes his bridge blog) but apparently they didn't want me.

Recently, I've been re-reading (for about the fourth time) Hugh Kelsey's truly excellent book on matchpoints. When you find yourself in a contract that appears to be a minority choice, you have to carefully consider what might happen at the other tables. This I've resolved to do.

Anyway, here's a hand that came up in the middle of a BBO Robot tournament, which I was winning at the time with 60-something percent.

What would you bid? I reasoned that I only needed one trick (and some sort of diamond fit) from partner to make 3NT. Of course, I only have one heart stopper and if partner's trick is slow then I'll probably go down. But surely that's a better plan than simply bidding 4D?

Now what? I could play a club up to the queen in dummy. If LHO had the ace, he'd might play low and I would have my 9 tricks. Some human defenders might fall for that. Should I consider simply conceding down one? No, I decided. There would be company in my contract and I should make every effort to take nine tricks. I decided on the club play, therefore. Unfortunately, my robot LHO was not brooking any funny business. The result?

I went down 7 tricks for -700 and a big round zero. Was I right in my strategy?

Most of my counterparts would probably take the safe 4D route, in which case they would all be going +130. Could there be any 3H contracts their way? Probably not, but if there were, they would make 140 or 170 depending on whether they had a diamond loser or not. Taking eight tricks, therefore would beat those pairs (if any). If half of the other pairs were in 4D, then conceding defeat would give me at best an average score and more likely average minus, let's say 30% for the sake of argument. If I was to sneak a club through and actually make my contract, I would be improving my score by 70% while risking my "safe" 30%.

However, I should have thought more about that all-important first trick! Having started with QJ9xxx and having decided to lead fourth best and seeing three hearts in dummy and partner, who raised, go in with the king, even the smallest brain would have known that the hearts were now running. He didn't need a lead from partner through the ten. Maybe if LHO had led the Q, seen two hearts in dummy, partner overtake with the K, he might have been worried that I had ATx in which case he would hope his partner had the CK.

In fact, there were nine +130s, a -100 (the one other playing 3NT), a -200 (defending a 3H contract), and, surprisingly, a +120 from 1NT making two. Playing it safe for -100 would have resulted in a 12.5% board. Needless to say, I didn't hold my first place. I dropped to third.

Sleepless in Seattle (conclusion)

It's not often that you pick up a ten-card suit. My partner was dealt this beauty in a side game at the Seattle NABC: ♠AKJT987542 ♥– ♦873 ♣–. We were at favorable vulnerability and my LHO opened with a weak notrump. Partner bid 4♠ and that ended the auction. I produced quite a useful dummy as the ♣J was led: ♠– ♥KT97 ♦A954 ♣AQ872. Incidentally, do you think I should have raised or made some try for slam?

Now, the question is how to make the greatest number of tricks. There isn't much you can do about trumps. You can't take a finesse and they either split or they don't. And in the latter case, there's no possible way to pull off a trump coup given that you would need to ruff seven cards in your hand and get back to dummy yet again! So, are there any realistic chances for an extra trick in a side suit? A singleton ace of hearts seems a bit unlikely given that there are nine out against you and RHO opened with 1NT (suggesting at least two and fewer than six). What about diamonds? They might be three-three but even then, you have nothing to pitch on the thirteener. No, the only realistic chance is in clubs. How realistic? Given that you only have one outside entry to dummy, you are going to need the king to fall doubleton. Is LHO really likely to have led J from KJ doubleton? I don't think so either. Therefore, you need RHO to hold Kx. You weren't thinking of covering with the queen, were you?

When you rise with the ♣A, the 9 drops on your right.  There are a total of eight clubs out and the a priori probability of righty having precisely K9 doubleton is only 0.3%! Is it even worth bothering with?  You betcha!  The K is doubleton.

However, only three of 11 declarers actually made 12 tricks, not including my very experienced partner, unfortunately.  Of course, some of these others might not have been given helpful club leads.  There was also a -350 and a -150 our way.  I imagine that these were uncompleted transfers or, more likely, an ill-fated attempt at 4NT or 6NT.

To conclude my commentary on Seattle, I will simply observe that this tiny probability (0.3%) yielded an additional 46% of the matchpoints on the board.  Well worth trying for, especially given that there was negligible down side to the play.

A couple of difficult hands from Auburn

In an otherwise decent effort Kim and I had a couple of tricky hands at the CMBA sectional in Auburn, MA.  First, a defensive problem.  Your hand is ♠– ♥JT864 ♦KQ ♣Q87654.  None vulnerable and partner deals and opens 1♠.  You bid a forcing 1NT.  LHO, a player who has never met a hand on which she could not find an overcall, bids 2♦.  Partner passes and it comes back around to you.  Double is primarily for take-out (but with the expectation after the pass that it might well be converted to penalty).  Bidding 2♥ and 3♣ both seem somewhat flawed.  So, let's say you do double and partner leaves it in.  Game for us seems unlikely, so 300 would be a top and even 100 might get most of the matchpoints.  In any case, you have to defend assuming that we are in the right contract.

Partner leads the ♠A (Ace from AK) and dummy comes down with an undeserved trick for declarer: ♠T8642 ♥9752 ♦3 ♣AT2.  Dummy follows low and you are at the cross-roads.  Partner won't be expecting your hand, that's for sure.  Maybe something like ♠93 ♥AT84 ♦72 ♣KJ765 or maybe ♠9 ♥AT864 ♦Q2 ♣K8765.  If you had one or two small trumps, you'd like to ruff a spade early so somehow you'd like to persuade partner to play a small one, if any, before your trumps get drawn.  You might do this by playing a low club then a low heart.  On the other hand, with your actual hand, you don't particularly want to waste any trumps on ruffing partner's losers.  Rather, you want partner to get dummy's entry off the table before the high spade could become good.  So, although this might typically suggest you have the king, I think the right card at trick one is the ♣8.

Unfortunately, neither of us defended optimally and on this occasion, declarer's hand was just good enough to take advantage and score 8 tricks for 180.  This wasn't an absolute bottom for us, but it was a low score.  Actually, it turns out that the normal contract was 3♣ by our side making exactly, so even +100 would not have been a good matchpoint score.

Here's a poor result that was entirely my fault, but is interesting theoretically, nonetheless.  I picked up ♠AK7 ♥98 ♦Q542 ♣KT85 in fourth seat at favorable vulnerability. Partner opened 1♥ and I responded a forcing 1 notrump.  Partner now rebid 2♥ which, in our system practically guarantees six pieces and tends to show a minimum hand strength-wise.  Obviously, I was going to bid game, but which game?  I felt that it might be advantageous to have the lead come up to my hand, especially on a minor suit lead, and bid 3NT – but I neglected three important factors.

First of all, partner's hand might be short on entries given the auction (or alternatively have a poor heart suit).  Both of these factors argue in favor of playing in a major suit game.  Secondly, the choice of notrump versus a major suit tends to work better with a 5-3 fit rather than a 6-2 fit.  Finally, choosing notrump over any 8-card major suit fit should generally only be considered with a plethora of high-card points, something like in the range 27-30.

So, to my contract of 3NT, a fourth-best deuce of spades was led.  Dummy came down much as expected with
♠J6 ♥AKJ764 ♦K8 ♣J97.  Obviously, I was going to try the J.  If it held, my judgment would be vindicated and I would likely make the same number of tricks as the heart declarers.  Unfortunately, the J was covered by the Q and I won with the Ace.  Now, I was definitely behind the heart declarers.  Any lead from the other defender would have likely given away a trick.  Not only that but I now had to be quite careful.  If the ♦K proved not to be an entry, it would be highly embarrassing to leave several hearts stranded in the dummy.

So, I turned to a couple of guidelines.  One was that if hearts were 3-2 I was destined to score badly.  The heart declarers would always score 20 points better than me.  If hearts were 4-1 offside, I'd be just as badly off, probably even worse.  That didn't bear thinking about.  But what if hearts were 4-1 on-side?  The heart declarers would all likely finesse the J and then try to drop the Q or T.  A first-round finesse was obviously called for, but which finesse?

That's when I turned to my Principle of Least Commitment for guidance.  This is the lazy man's way of avoiding having to learn all 656 suit combinations from the Bridge Encyclopedia.  In this case, least commitment suggests running the 9.  The advantage of running the 9 is that if RHO wins with the Q, you know where the T is (unless RHO is very devious indeed).  If you run the 9 and it loses to the T of course, you know nothing about the Q and if you finesse the J and it loses to the Q, you know nothing of the T. 

If entries to dummy were not a problem (or if hearts were trumps), then the best play is to cash a high heart, cross over and finesse the J.  You'll make 6 tricks 37% of the time and 5 tricks 88% of the time.  But if we assume no outside entry to dummy, then we essentially want to duck a trick to maintain our link.  Again, this suggests running the 9, which is what I did.  It lost to the T.

Another way of looking at it is this: if indeed there is no further entry to dummy, running the 9 first will result in either 2 or 5 tricks in the suit, assuming that the hearts are distributed unfavorably: Q532–T or T532–Q.  Finessing the J first will result in either 3 or 2 tricks.

A spade came back and now I had another decision to make.   So far, my strategy was not panning out.  The heart declarers would finesse the J and see the T come up on their left.  Then they'd bang down the top hearts and hope to drop the Q.  If that happened, I'd lose.  Was there a way to win?  Yes: take another finesse in hearts.  But wait!  If that lost to the Q, I might be in the ignominious position of not taking any heart tricks at all and going down quite a few.

Here's where I goofed.  I got scared.  I didn't "stay with the program."  I couldn't bear the thought of looking so silly so I played off the A and K.  The Q failed to appear.  She was exactly where I needed her for a good board.  What an idiot!  I ended up with -50 while my competition were all +420.  I might still have ended up with 400 which would have been good for slightly over average.

Matchpoint Bridge

Recently, I've been re-reading what surely must be the best book that deals specifically with playing matchpoints, and one of the best bridge books of all time.  I refer of course to Hugh Kelsey's Match-point Bridge (ISBN 0 571 11437 7).  After the excellent introduction, there are five parts, each of which is treated in three different sub-parts: Bidding, Dummy play and Defence.  The five parts are: Fundamentals, Contesting the Part Score, The Lead, Sacrifice and Deception.

In the introduction he makes the very important point, obvious if you stop to think about it -- but not so clear otherwise -- that the two phases of the matchpoint game, the auction and the play are inextricably linked.  If your side has fallen short in the auction, then you are destined to a poor score so that explanation of the auction must be incorrect.  You have to play on the assumption that you actually did get the auction right.  An example which we've all experienced is to have missed an "obvious" game which looks likely to make.  Clearly, you are headed for a bad board -- unless that is the game goes down.  You should therefore assume that the cards are wrong for the game and right for you.

Here's one of my favorite sections from "Sacrifice - Dummy Play."

♠K7 ♥K4 ♦T763 ♣J9643 ♠AQT863 ♥63 ♦A52 ♣85

None Vulnerable, Dealer West West North East South 1♥ p 2♥ 2♠ 4♥ 4♠ p p Double p p p

West leads the King of diamonds against your doubled contract of four spades and you win with the ace. How do you plan the play?

The situation looks grim. Partner's raise was eccentric, to say the least, and it is a safe bet that most of the other North-South pairs will choose to defend against four hearts. In order to restrict the penalty to 300 in four spades doubled, you will need to make six trump tricks and the king of hearts in addition to the diamond already in the bag. But if the heart ace is right for you, it is wrong for the opponents -- that is to say the heart king would have scored in defence. It follows that with a normal spade break you would have four defensive tricks against a heart contract, and the pairs defending against four hearts will register plus scores. That is a possibility which must not be entertained. You must assume that your idiot-partner did the right thing in bidding four spades and the only distribution to make that possible is a 4-1 trump break with the singleton in the West hand. Accordingly, you should lead a spade to the king and if the knave does not appear, finesse the ten on the way back.
The book is full of such gems, not all quite as humorous as this one, but each equally brilliant in its analysis and good advice.